Chore: lazy loading for rules

[mysticatea]

What is the purpose of this pull request? (put an "X" next to item)

[X] Other, please explain: improve performance.

What changes did you make? (Give an overview)

This PR improves performance by avoiding to load unused rules.
But this PR keeps static linking for bundlers can recognize rules.

• lib/util/lazy-loading-rule-map.js ... The map of built-in rules. This map will load each rule at the first access.
• lib/cli.js ... This file will iterate all rules at the first access of rulesMeta.
• lib/testers/rule-tester.js ... This file no longer iterates all rules.
• conf/eslint-recommended.js ... This file no longer iterates all rules.
• Makefile.js ... The test verifies if conf/eslint-recommended.js is correct.

---

Impacts:

eslint lib/linter.js --no-eslintrc --config conf/eslint-recommended.js --parser-options ecmaVersion:2018
# 0.73s → 0.56s (23% speed up)

eslint lib/linter.js
# 1.25s → 1.20s (4% speed up)

mocha "tests/{bin,lib,tools}/**/*.js" --reporter progress
# 52s → 49s (5% speed up)

Especially, it has a lerge impact if it used fewer rules such as eslint:recommended.

---

npm run perf:

# BEFORE -----------------------------------------------------------------------
Loading:
  Load performance Run #1:  421.4627ms
  Load performance Run #2:  449.755299ms
  Load performance Run #4:  460.1405ms
  Load performance Run #5:  453.075ms
  Load Performance median:  453.075ms


Single File:
  CPU Speed is 3192 with multiplier 13000000
  Performance Run #1:  3998.2129ms
  Performance Run #2:  3855.885199ms
  Performance Run #3:  3704.6185ms
  Performance Run #4:  3810.938001ms
  Performance Run #5:  3731.5751ms

  Performance budget ok:  3810.938001ms (limit: 4072.6817042606517ms)


Multi Files (0 files):
  CPU Speed is 3192 with multiplier 39000000
  Performance Run #1:  11271.758299ms
  Performance Run #2:  10577.971301ms
  Performance Run #3:  10542.4316ms
  Performance Run #4:  10947.606701ms
  Performance Run #5:  10122.6511ms

  Performance budget ok:  10577.971301ms (limit: 12218.045112781954ms)

# AFTER ------------------------------------------------------------------------
Loading:
  Load performance Run #1:  220.293599ms
  Load performance Run #2:  215.602099ms
  Load performance Run #3:  210.9296ms
  Load performance Run #4:  222.679ms
  Load performance Run #5:  217.569701ms

  Load Performance median:  217.569701ms


Single File:
  CPU Speed is 3192 with multiplier 13000000
  Performance Run #1:  3839.186ms
  Performance Run #2:  3770.319299ms
  Performance Run #3:  3685.686001ms
  Performance Run #4:  3779.8463ms
  Performance Run #5:  3650.4698ms

  Performance budget ok:  3770.319299ms (limit: 4072.6817042606517ms)


Multi Files (0 files):
  CPU Speed is 3192 with multiplier 39000000
  Performance Run #1:  10705.549801ms
  Performance Run #2:  10394.445ms
  Performance Run #3:  10542.7665ms
  Performance Run #4:  10355.708099ms
  Performance Run #5:  10419.626301ms

  Performance budget ok:  10419.626301ms (limit: 12218.045112781954ms)

• Load performance 52% speed up.
• Single File 4% speed up.
• Multi Files 1% speed up.

Is there anything you'd like reviewers to focus on?

Can we go to this direction?